论文部分内容阅读
在光电跟踪系统目标状态预测过程中,噪声统计特性不确定是导致滤波精度下降的主要原因之一。针对该问题,提出了一种基于“当前”统计模型的扩展H∞滤波(EHF)方法,并与扩展卡尔曼(EKF)滤波方法做了仿真比较。仿真结果证明,在白噪声条件下,EKF滤波方法虽然精度略高,但收敛速度要差于EHF方法;在有色噪声条件下,EKF滤波方法几乎不能完成滤波,而EHF滤波器可以保持和白噪声假设条件相当的滤波精度,并且在目标发生机动的情况下该方法仍具有较强的鲁棒性。
In the process of predicting the target state of photoelectric tracking system, the uncertainty of the noise statistics is one of the main reasons for the decrease of the filtering accuracy. To solve this problem, an extended H∞ filter (EHF) method based on the “current” statistical model is proposed and compared with the extended Kalman filtering method. The simulation results show that the EKF filtering method is less accurate than the EHF method under the white noise condition, although the EKF filtering method is slightly more accurate than the EHF method. In the case of colored noise, the EKF filtering method can hardly achieve the filtering, while the EHF filter can maintain the white noise It is assumed that the conditions are equivalent to the filtering accuracy, and the method still has strong robustness in the case of maneuvering of the target.